Mindfulness Training as a Clinical Intervention: A Conceptual and Empirical Review

Page created by Joann Todd
 
CONTINUE READING
Mindfulness Training as a Clinical Intervention:
A Conceptual and Empirical Review
Ruth A. Baer, University of Kentucky

Interventions based on training in mindfulness skills are       (Goleman & Schwartz, 1976; Kabat-Zinn, 1982). The cur-
becoming increasingly popular. Mindfulness involves in-         rent mindfulness literature describes numerous meditation
tentionally bringing one’s attention to the internal and ex-    exercises designed to develop mindfulness skills (Hanh,
ternal experiences occurring in the present moment, and         1976; Kabat-Zinn, 1990, 1994; Linehan, 1993b). Many
is often taught through a variety of meditation exercises.      encourage individuals to attend to the internal experiences
This review summarizes conceptual approaches to mind-
                                                                occurring in each moment, such as bodily sensations,
                                                                thoughts, and emotions. Others encourage attention to
fulness and empirical research on the utility of mindfulness-
                                                                aspects of the environment, such as sights and sounds
based interventions. Meta-analytic techniques were in-
                                                                (Kabat-Zinn, 1994; Linehan, 1993b). All suggest that
corporated to facilitate quantification of findings and
                                                                mindfulness should be practiced with an attitude of non-
comparison across studies. Although the current empir-
                                                                judgmental acceptance. That is, phenomena that enter the
ical literature includes many methodological flaws, find-       individual’s awareness during mindfulness practice, such as
ings suggest that mindfulness-based interventions may           perceptions, cognitions, emotions, or sensations, are ob-
be helpful in the treatment of several disorders. Method-       served carefully but are not evaluated as good or bad, true
ologically sound investigations are recommended in order        or false, healthy or sick, or important or trivial (Marlatt &
to clarify the utility of these interventions.                  Kristeller, 1999). Thus, mindfulness is the nonjudgmental
   Key words: mindfulness, meditation, meta-analysis,           observation of the ongoing stream of internal and external
treatment outcome. [Clin Psychol Sci Prac 10: 125–143,          stimuli as they arise.
2003]                                                               Until recently, mindfulness has been a relatively unfa-
                                                                miliar concept in much of our culture (Kabat-Zinn, 1982),
Mindfulness is a way of paying attention that originated in     perhaps because of its origins in Buddhism. Kabat-Zinn
Eastern meditation practices. It has been described as          (2000) suggests that mindfulness practice may be beneficial
“bringing one’s complete attention to the present experi-       to many people in Western society who might be un-
ence on a moment-to-moment basis” (Marlatt & Kris-              willing to adopt Buddhist traditions or vocabulary. Thus,
teller, 1999, p. 68) and as “paying attention in a particular   Western researchers and clinicians who have introduced
way: on purpose, in the present moment, and nonjudg-            mindfulness practice into mental health treatment pro-
mentally” (Kabat-Zinn, 1994, p. 4). The ability to direct       grams usually teach these skills independently of the reli-
one’s attention in this way can be developed through the        gious and cultural traditions of their origins (Kabat-Zinn,
practice of meditation, which is defined as the intentional      1982; Linehan, 1993b). In the current empirical literature,
self-regulation of attention from moment to moment              clinical interventions based on training in mindfulness skills
                                                                are described with increasing frequency, and their popu-
                                                                larity appears to be growing rapidly. According to Salmon,
Send correspondence to Ruth A. Baer, Department of Psychol-     Santorelli, and Kabat-Zinn (1998), over 240 hospitals and
ogy, 115 Kastle Hall, University of Kentucky, Lexington, KY     clinics in the United States and abroad were offering stress
40506-0044. E-mail: rbaer@uky.edu.                              reduction programs based on mindfulness training as of

DOI: 10.1093/clipsy/bpg015
 2003 AMERICAN PSYCHOLOGICAL ASSOCIATION D12                                                                             125
1997. Mindfulness training is also a central component of        described in this review often are directed toward the in-
dialectical behavior therapy (Linehan, 1993a, 1993b), an         ner experiences of the individual (e.g., thoughts, emo-
increasingly popular approach to the treatment of border-        tions) and emphasize a less goal-directed, nonjudgmental
line personality disorder. The empirical literature on the       observation. Langer (1989) has cautioned against drawing
effects of mindfulness training contains many method-             unwarranted parallels between the two forms of mindful-
ological weaknesses, but it suggests that mindfulness inter-     ness, noting that they are derived from very different his-
ventions may lead to reductions in a variety of problematic      torical and cultural backgrounds.
conditions, including pain, stress, anxiety, depressive re-
lapse, and disordered eating (e.g., Kabat-Zinn, 1982;Kabat-      INTERVENTIONS BASED ON MINDFULNESS TRAINING
Zinn et al., 1992; Kristeller & Hallett, 1999; Shapiro,          Mindfulness-Based Stress Reduction
Schwartz, & Bonner, 1998; Teasdale et al., 2000).                The most frequently cited method of mindfulness training
    This review summarizes the recent literature on mind-        is the mindfulness-based stress reduction (MBSR) pro-
fulness training as a clinical intervention. First, current      gram, formerly known as the stress reduction and relax-
methods for teaching mindfulness skills are described.           ation program (SR-RP; Kabat-Zinn, 1982, 1990). It was
Next, conceptual approaches that articulate how mindful-         developed in a behavioral medicine setting for populations
ness skills may be helpful in treating clinical conditions are   with a wide range of chronic pain and stress-related disor-
summarized. Finally, the empirical literature on the effects      ders. The program is conducted as an 8- to 10-week course
of mindfulness training is reviewed.                             for groups of up to 30 participants who meet weekly for
    This review does not address transcendental meditation       2–2.5 hr for instruction and practice in mindfulness med-
(TM) and other concentration-based approaches, which             itation skills, together with discussion of stress, coping, and
have been reviewed elsewhere (Delmonte, 1985; Smith,             homework assignments. An all-day (7–8-hr) intensive
1975). Concentration-based approaches train participants         mindfulness session usually is held around the sixth week.
to restrict the focus of attention to a single stimulus, such    Several mindfulness meditation skills are taught. For ex-
as a word (e.g., a mantra), sound, object, or sensation.         ample, the body scan is a 45-min exercise in which atten-
When attention wanders, it is redirected to the object of        tion is directed sequentially to numerous areas of the body
meditation. No attention is paid to the nature of the dis-       while the participant is lying down with eyes closed. Sen-
traction. Mindfulness meditation, in contrast, involves ob-      sations in each area are carefully observed. In sitting med-
servation of constantly changing internal and external           itation, participants are instructed to sit in a relaxed and
stimuli as they arise.                                           wakeful posture with eyes closed and to direct attention to
    This review also does not address Langer’s (1989, 1997)      the sensations of breathing. Hatha yoga postures are used
cognitive model of mindfulness, which includes alertness         to teach mindfulness of bodily sensations during gentle
to distinctions, context, and multiple perspectives, open-       movements and stretching. Participants also practice mind-
ness to novelty, and orientation in the present (Sternberg,      fulness during ordinary activities like walking, standing,
2000). Mindfulness interventions studied by Langer and           and eating.
colleagues (e.g., Langer, 1989; Langer & Moldoveanu,                 Participants in MBSR are instructed to practice these
2000) often include teaching participants to consider in-        skills outside group meetings for at least 45 min per day, six
formation or situations from multiple perspectives or            days per week. Audiotapes are used early in treatment, but
within new contexts in order to increase learning or cre-        participants are encouraged to practice without tapes after
ativity. Although this concept of mindfulness shares with        a few weeks. For all mindfulness exercises, participants are
meditative approaches an emphasis on flexible awareness in        instructed to focus attention on the target of observation
the present, several important differences can be noted.          (e.g., breathing or walking) and to be aware of it in each
Langer’s (1989) mindfulness interventions usually involve        moment. When emotions, sensations, or cognitions arise,
working with material external to the participants, such as      they are observed nonjudgmentally. When the participant
information to be learned or manipulated, and often in-          notices that the mind has wandered into thoughts, mem-
clude active, goal-oriented cognitive tasks, such as solving     ories, or fantasies, the nature or content of them is briefly
problems. In contrast, the meditation-based approaches           noted, if possible, and then attention is returned to the

CLINICAL PSYCHOLOGY: SCIENCE AND PRACTICE             •   V10 N2, SUMMER 2003                                              126
present moment. Thus, participants are instructed to no-        view, which postulates that reality consists of opposing
tice their thoughts and feelings but not to become absorbed     forces. The synthesis of these forces leads to a new reality,
in their content (Kabat-Zinn, 1982). Even judgmental            which in turn consists of opposing forces, in a continual
thoughts (e.g., “this is a foolish waste of time”) are to be    process of change. In DBT, the most central dialectic is the
observed nonjudgmentally. Upon noticing such a thought,         relationship between acceptance and change. Clients are
the participant might label it as a judgmental thought, or      encouraged to accept themselves, their histories, and their
simply as “thinking,” and then return attention to the pres-    current situations exactly as they are, while working in-
ent moment. An important consequence of mindfulness             tensively to change their behaviors and environments in
practice is the realization that most sensations, thoughts,     order to build a better life. The synthesis of this apparent
and emotions fluctuate, or are transient, passing by “like       contradiction is a central goal of DBT.
waves in the sea” (Linehan, 1993b, p. 87).                          DBT includes a wide range of cognitive and behavioral
                                                                treatment procedures, most of which are designed to
Mindfulness-Based Cognitive Therapy                             change thoughts, emotions, or behaviors. Mindfulness
Teasdale, Segal, and Williams (1995) proposed that the          skills are taught in DBT within the context of synthesizing
skills of attentional control taught in mindfulness medita-     acceptance and change. Although the skills taught are sim-
tion could be helpful in preventing relapse of major de-        ilar to those targeted in MBSR, including nonjudgmental
pressive episodes. Their information-processing theory of       observation of thoughts, emotions, sensations, and envi-
depressive relapse suggests that individuals who have ex-       ronmental stimuli, the concepts are organized somewhat
perienced major depressive episodes are vulnerable to re-       differently. For example, Linehan (1993a, 1993b) describes
currences whenever mild dysphoric states are encountered,       three mindfulness “what” skills (observe, describe, par-
because these states may reactivate the depressive thinking     ticipate) and three mindfulness “how” skills (nonjudg-
patterns present during the previous episode, or episodes,      mentally, one-mindfully, effectively). DBT clients learn
thus precipitating a new episode. Mindfulness-based cog-        mindfulness skills in a year-long weekly skills group, which
nitive therapy (MBCT) is a manualized (Segal, Williams, &       also covers interpersonal effectiveness, emotion regulation,
Teasdale, 2002) 8-week group intervention based largely on      and distress tolerance skills. Clients work with their indi-
Kabat-Zinn’s (1990) MBSR program. It incorporates ele-          vidual therapists on applying skills learned in group to their
ments of cognitive therapy that facilitate a detached or de-    daily lives.
centered view of one’s thoughts, including statements such          Linehan (1994) notes that some severely impaired in-
as “thoughts are not facts” and “I am not my thoughts.”         dividuals may be unable or unwilling to meditate as
This decentered approach also is applied to emotions and        extensively as Kabat-Zinn’s (1990) MBSR program rec-
bodily sensations. MBCT is designed to prevent depressive       ommends. Thus, DBT does not prescribe a specific fre-
relapse by teaching formerly depressed individuals to ob-       quency or duration of mindfulness practice. Instead, goals
serve their thoughts and feelings nonjudgmentally, and to       for mindfulness practice are established by individual
view them simply as mental events that come and go,             clients and their therapists. DBT offers numerous mind-
rather than as aspects of themselves, or as necessarily accu-   fulness exercises from which clients may choose (some
rate reflections of reality. This attitude toward depression-    adapted from Hanh, 1976). In one example, clients imag-
related cognitions is believed to prevent the escalation of     ine that the mind is a conveyor belt. Thoughts, feelings,
negative thoughts into ruminative patterns (Teasdale et al.,    and sensations that come down the belt are observed, la-
1995).                                                          beled, and categorized. In another exercise, clients imag-
                                                                ine that the mind is the sky, and that thoughts, feelings, and
INTERVENTIONS INCORPORATING                                     sensations are clouds that they watch passing by. Several
MINDFULNESS TRAINING                                            variations on observing the breath are taught, including
Dialectical Behavior Therapy                                    following the breath in and out, counting breaths, coordi-
Dialectical behavior therapy (DBT) is a multifaceted ap-        nating breathing with footsteps while walking, and fol-
proach to the treatment of borderline personality disorder      lowing the breath while listening to music. Some exercises
(Linehan, 1993a, 1993b). It is based on a dialectical world-    encourage mindful awareness during everyday activities,

MINDFULNESS TRAINING           •   BAER                                                                                   127
such as making tea, washing dishes or clothes, cleaning         eliminated. Instead, urges must be accepted as normal re-
house, or taking a bath.                                        sponses to appetitive cues. Mindfulness skills enable the
                                                                client to observe the urges as they appear, accept them non-
Acceptance and Commitment Therapy                               judgmentally, and cope with them in adaptive ways.
Acceptance and commitment therapy (ACT; Hayes,
Strosahl, & Wilson, 1999) is theoretically based in con-        CONCEPTUAL APPROACHES: HOW MINDFULNESS
temporary behavior analysis (Hayes & Wilson, 1993). Al-         SKILLS MAY HELP
though ACT does not describe its treatment methods in           The authors of these treatment strategies have suggested
terms of mindfulness or meditation, it is included here be-     several mechanisms that may explain how mindfulness
cause several of its strategies are consistent with the mind-   skills can lead to symptom reduction and behavior change.
fulness approaches described. Clients in ACT are taught to
recognize an observing self who is capable of watching his      Exposure
or her own bodily sensations, thoughts, and emotions.           The first published study of the effects of MBSR (Kabat-
They are encouraged to see these phenomena as separate          Zinn, 1982) described its application in patients with
from the person having them. For example, they are taught       chronic pain. MBSR is based, in part, on traditional med-
to say, “I’m having the thought that I’m a bad person,”         itation practices, which often include extended periods of
rather than “I’m a bad person” (Kohlenberg, Hayes, &            motionless sitting. Although a relaxed posture typically is
Tsai, 1993, p. 588). They also are encouraged to experi-        adopted, prolonged motionlessness can lead to pain in
ence thoughts and emotions as they arise, without judging,      muscles and joints. Mindfulness meditation instructors of-
evaluating, or attempting to change or avoid them. Hayes        ten encourage students not to shift position to relieve the
(1987) describes an exercise in which the client imagines       pain, but instead to focus careful attention directly on the
that his or her thoughts are written on signs carried by        pain sensations, and to assume a nonjudgmental attitude
parading soldiers. The client’s task is to observe the parade   toward these sensations, as well as toward the various cog-
of thoughts without becoming absorbed in any of them.           nitions (“this is unbearable”) emotions (anxiety, anger),
ACT explicitly teaches clients to abandon attempts to con-      and urges (to shift position) that often accompany pain
trol thoughts and feelings, but instead to observe them non-    sensations. The ability to observe pain sensations non-
judgmentally and accept them as they are, while changing        judgmentally is believed to reduce the distress associated
their behaviors in constructive ways to improve their lives     with pain.
(Hayes, 1994).                                                      Kabat-Zinn (1982) suggests that application of this
                                                                strategy by chronic pain patients might serve several func-
Relapse Prevention                                              tions. For example, prolonged exposure to the sensations
Relapse prevention (RP; Marlatt & Gordon, 1985) is a            of chronic pain, in the absence of catastrophic conse-
cognitive-behavioral treatment package designed to fore-        quences, might lead to desensitization, with a reduction
stall relapses in individuals treated for substance abuse.      over time in the emotional responses elicited by the pain
Mindfulness skills are included as a technique for coping       sensations. Thus, the practice of mindfulness skills could
with urges to engage in substance use. Marlatt (1994) notes     lead to the ability to experience pain sensations without
that mindfulness involves acceptance of the constantly          excessive emotional reactivity. Even if pain sensations were
changing experiences of the present moment, whereas ad-         not reduced, suffering and distress might be alleviated.
diction is an inability to accept the present moment and a          Kabat-Zinn et al. (1992) describe a similar mechanism
persistent seeking of the next “high” associated with the       for the potential effects of mindfulness training on anxi-
addiction. The metaphor of “urge surfing” encourages             ety and panic. Sustained, nonjudgmental observation of
clients to imagine that urges are ocean waves that grow         anxiety-related sensations, without attempts to escape or
gradually until they crest and subside. The client “rides”      avoid them, may lead to reductions in the emotional reac-
the waves without giving in to the urges, thus learning         tivity typically elicited by anxiety symptoms. This approach
that urges will pass. However, the client also learns that      is similar to the interoceptive exposure strategy described
new urges will appear and that these urges cannot easily be     by Barlow and Craske (2000), who instruct clients to in-

CLINICAL PSYCHOLOGY: SCIENCE AND PRACTICE            •   V10 N2, SUMMER 2003                                            128
duce symptoms of panic through exercises such as hyper-          decentered view of one’s thoughts encouraged by mind-
ventilation and aerobic activity, and to practice tolerating     fulness training may interfere with ruminative patterns
these sensations until they subside. In contrast, however,       believed to be characteristic of depressive episodes (Nolen-
mindfulness training does not include the deliberate in-         Hoeksema, 1991). That is, mindfulness training may en-
duction of panic symptoms. Instead, participants are in-         able formerly depressed individuals to notice depressogenic
structed to observe these sensations nonjudgmentally when        thoughts and to redirect attention to other aspects of the
they naturally arise.                                            present moment, such as breathing, walking, or environ-
    Linehan (1993a, 1993b) describes individuals with bor-       mental sounds, thus avoiding rumination. Teasdale has
derline personality disorder as emotion phobic. That is,         described this perspective on one’s thoughts as “meta-
they are often afraid of experiencing strong negative affec-      cognitive insight.” Teasdale et al. (1995) also note that a
tive states. This fear is understandable, because their neg-     practical advantage of mindfulness skills in encouraging
ative affective states typically are very intense. However,       cognitive change is that they can be practiced at any time,
their attempts to avoid these states often have maladaptive      including during periods of remission, when depresso-
consequences. Linehan (1993a, 1993b) suggests that pro-          genic thinking may be occurring too rarely to permit reg-
longed observation of current thoughts and emotions,             ular practice of traditional cognitive therapy exercises, such
without trying to avoid or escape them, can be seen as an        as identification and disputing of cognitive distortions.
example of exposure, which should encourage the extinc-          That is, a mindful perspective about one’s thoughts can be
tion of fear responses and avoidance behaviors previously        applied to all thoughts.
elicited by these stimuli. Thus, the practice of mindfulness
skills may improve patients’ ability to tolerate negative        Self-Management
emotional states and ability to cope with them effectively.       Several authors have noted that improved self-observation
                                                                 resulting from mindfulness training may promote use of a
Cognitive Change                                                 range of coping skills. For example, Kabat-Zinn (1982)
Several authors have noted that the practice of mindfulness      suggests that increased awareness of pain sensations and
may lead to changes in thought patterns, or in attitudes         stress responses as they occur may enable individuals to en-
about one’s thoughts. For example, Kabat-Zinn (1982,             gage in a variety of coping responses, including skills not
1990) suggests that nonjudgmental observation of pain and        included in their treatment program. Kristeller and Hallett
anxiety-related thoughts may lead to the understanding           (1999) suggest that the self-observation skills developed
that they are “just thoughts,” rather than reflections of truth   through mindfulness training might lead to improved
or reality, and do not necessitate escape or avoidance           recognition of satiety cues in binge eaters, as well as in-
behavior. Similarly, Linehan (1993a, 1993b) notes that ob-       creased ability to observe urges to binge without yielding
serving one’s thoughts and feelings and applying descrip-        to them. Marlatt (1994) suggests a similar effect in patients
tive labels to them encourages the understanding that they       recovering from addictions. Teasdale et al. (1995) note that
are not always accurate reflections of reality. For example,      mindfulness training encourages awareness of all cognitive
feeling afraid does not necessarily mean that danger is im-      and emotional events as they occur, including those that
minent, and thinking “I am a failure” does not make it           may be early signs of potential depressive relapse. Thus,
true. Kristeller and Hallett (1999), in a study of MBSR in       mindfulness training may promote recognition of early
patients with binge eating disorder, cite Heatherton and         signs of a problem, at a time when application of previously
Baumeister’s (1991) theory of binge eating as an escape          learned skills will be most likely to be effective in prevent-
from self-awareness and suggest that mindfulness training        ing the problem. Linehan (1993b) suggests that nonjudg-
might develop nonjudgmental acceptance of the aversive           mental observation and description permits recognition
cognitions that binge-eaters are thought to be avoiding,         of the consequences of behaviors (e.g., irritating one’s boss
such as unfavorable comparisons of self to others and per-       with frequent lateness) in place of global judgments about
ceived inability to meet others’ demands.                        the self (e.g., “I am a bad employee”). This recognition
   Teasdale (1999) and Teasdale et al. (1995), in their          may lead to more effective behavior change, including re-
discussion of MBCT, suggest that the nonjudgmental,              duction of impulsive, maladaptive behaviors. Linehan

MINDFULNESS TRAINING        •   BAER                                                                                       129
(1993b) also suggests that learning to focus “one-             erated, rather than fearsome and dangerous experiences to
mindfully” on the present moment develops control of at-       be avoided, even at the cost of significant maladaptive be-
tention, a useful skill for individuals who have difficulty      havior.
completing important tasks because they are distracted by          All of the treatment programs reviewed here include ac-
worries, memories, or negative moods.                          ceptance of pain, thoughts, feelings, urges, or other bod-
                                                               ily, cognitive, and emotional phenomena, without trying
Relaxation                                                     to change, escape, or avoid them. Kabat-Zinn (1990) de-
The relationship between meditation and relaxation is          scribes acceptance as one of several foundations of mind-
somewhat complex. Several authors (Goldenberg et al.,          fulness practice. DBT provides explicit training in several
1994; Kabat-Zinn et al., 1998; Kaplan, Goldenberg, &           mindfulness techniques designed to promote acceptance of
Galvin-Nadeau, 1993) have suggested that mindfulness-          reality. Thus, it appears that mindfulness training may pro-
based stress reduction may be applicable to stress-related     vide a method for teaching acceptance skills.
medical disorders, including psoriasis and fibromyalgia.
These authors note that meditation often induces relax-        Relationship Between Mindfulness Training and Cognitive-
ation, which may contribute to the management of these         Behavioral Approaches
disorders. The induction of relaxation through various         This discussion suggests that mindfulness training is con-
meditation strategies has been well documented (Benson,        sistent with cognitive-behavioral treatment procedures in
1975; Orme-Johnson, 1984; Wallace, Benson, & Wilson,           several ways. Training in self-directed attention can result
1984). However, the purpose of mindfulness training is         in sustained exposure to sensations, thoughts, and emo-
not to induce relaxation, but instead to teach nonjudg-        tions, resulting in desensitization of conditioned responses
mental observation of current conditions, which might          and reduction of avoidance behavior. Cognitive change
include autonomic arousal, racing thoughts, muscle ten-        appears to result from viewing one’s thoughts as temporary
sion, and other phenomena incompatible with relaxation.        phenomena without inherent worth or meaning, rather
In addition, evidence suggests that relaxation effects are      than as necessarily accurate reflections of reality, health,
not unique to meditation, but are common to many re-           adjustment, or worthiness. Practice of meditation also may
laxation strategies (Shapiro, 1982). Thus, although practice   lead to relaxation and improved self-management.
of mindfulness exercises may lead to relaxation, this out-         However, mindfulness training differs from traditional
come may not be a primary reason for engaging in mind-         cognitive-behavioral treatment in important ways. For ex-
fulness skills.                                                ample, mindfulness training does not include the evalua-
                                                               tion of thoughts as rational or distorted, or systematic
Acceptance                                                     attempts to change thoughts judged to be irrational. In-
The relationship between acceptance and change is a cen-       stead, participants are taught to observe their thoughts, to
tral concept in current discussions of psychotherapy (Hayes,   note their impermanence, and to refrain from evaluating
Jacobsen, Follette, & Dougher, 1994). Hayes (1994) sug-        them. Another important difference is that traditional
gests that acceptance involves “experiencing events fully      cognitive-behavioral procedures usually have a clear goal,
and without defense, as they are” (p. 30), and notes that      such as to change a behavior or thinking pattern. In con-
empirically oriented clinicians may have overemphasized        trast, mindfulness meditation is practiced with a seemingly
the importance of changing all unpleasant symptoms,            paradoxical attitude of nonstriving. That is, although a task
without recognizing the importance of acceptance. For          is prescribed (e.g., sit still, close your eyes, and pay atten-
example, an individual who experiences panic attacks may       tion), no specific goal is adopted. Participants are not to
engage in numerous maladaptive behaviors in an attempt         strive to relax, reduce their pain, or change their thoughts
to prevent future attacks, including drug and alcohol abuse,   or emotions, although they may have sought treatment for
avoidance of important activities, and excessive anxious       these purposes. They are simply to observe whatever is
vigilance toward bodily states. If the individual could ac-    happening in each moment without judging it. Finally,
cept that panic attacks may occasionally occur and that        mindfulness researchers have suggested that effective teach-
they are time-limited and not dangerous, panic attacks         ing of mindfulness skills by mental health professionals re-
would become unpleasant but brief experiences to be tol-       quires that they engage in their own regular mindfulness

CLINICAL PSYCHOLOGY: SCIENCE AND PRACTICE           •   V10 N2, SUMMER 2003                                               130
practice (Segal et al., 2002). Professionals using more tra-      sure, Mc = the mean of the comparison group on that mea-
ditional cognitive-behavioral strategies generally are not        sure, and SDp = the pooled standard deviation of the two
expected to engage in regular practice of the skills they are     groups. If means or standard deviations were not provided,
teaching.                                                         effect sizes were calculated from the significance level (p).
    Although the practice of mindfulness generally involves       For studies using within-groups designs, effect sizes were
acceptance of current reality, rather than systematic at-         calculated from t or F (1 df ), or from the significance level
tempts to change reality, individuals who practice these          when t or F were not reported. Calculations of effect sizes
skills may experience reductions in a variety of symptoms.        relied on methods described by Rosenthal (1984).
The empirical literature addressing this issue is reviewed
next.                                                             Overview of Treatment Literature
                                                                  Studies examining the effects of mindfulness-based inter-
EMPIRICAL RESEARCH ON MINDFULNESS-BASED                           ventions are summarized in Table 1. The studies are
INTERVENTIONS                                                     grouped by participant population, beginning with stud-
The empirical literature investigating the effects of mind-        ies of chronic pain patients. Next are studies of patients
fulness-based interventions is reviewed here. Meta-analytic       with other Axis I disorders (anxiety, eating, and major de-
procedures were incorporated to facilitate quantification of       pressive disorders), followed by studies of patients with
findings and comparisons across studies. To locate relevant        other medical problems (fibromyalgia, psoriasis, and can-
studies, a computer search (using PsycInfo and Medline            cer). Next are studies with mixed populations, including
databases) was conducted of articles and chapters including       psychotherapy and medical patients. The last group in-
the terms mindfulness or meditation. Reference lists of all       cludes studies of nonclinical populations (students and
articles were searched for additional articles. Studies were      other volunteers). Within each group studies are listed in
included if they were published in English and compared           order of publication date.
a group of participants trained in mindfulness with a group           Sample sizes in these studies have ranged from 16 to
not trained, or a group who provided data before and af-          142. Mean age of participants has ranged from 38–50 years,
ter mindfulness training. Unpublished dissertations and           with a mean of 45 years. Gender ratio of the samples has
conference papers were excluded, as were studies address-         ranged from 0 to 46% male. Education and race/ethnicity
ing nonmindfulness forms of meditation, such as transcen-         were rarely reported.
dental meditation (TM). Studies of mindfulness as defined              Nine studies used pre-post designs with no control
by Langer (1989, 1997) also were excluded, for reasons            group. Nine used between-groups designs with Treatment
described earlier. Twenty-one studies meeting these cri-          As Usual (TAU) or waiting-list control groups. Most stud-
teria were found.                                                 ies used the 8–10 week MBSR group intervention (Kabat-
    For each study, several demographic and methodolog-           Zinn,1982, 1990), or a variation of this intervention
ical variables were coded, including number, type, and            tailored to the population under study. Two studies (Teas-
characteristics of participants, research design, the nature of   dale et al., 2000; Williams, Teasdale, Segal, & Soulsby,
the mindfulness intervention, type of comparison group,           2000) examined MBCT. Dependent variables have in-
whether participants were randomly assigned to interven-          cluded a variety of self-report measures of pain, other med-
tion or comparison groups, the dependent variables re-            ical symptoms, anxiety, depression, eating behaviors, and
ported, and follow-up intervals and data.                         general psychological functioning, as well as objective
    Effect sizes (Cohen’s d) were calculated for all studies       measures such as analysis of urine chemistry.
that provided sufficient data. Cohen’s d expresses effect size           For each study effect sizes were calculated separately
in standard deviation units; thus, an effect size of 1.0 on a      for each dependent measure completed at the conclusion
given dependent measure indicates that the treatment              of treatment and at all reported follow-up intervals. Post-
group scored one standard deviation better, on average,           treatment effect sizes then were averaged across dependent
than the comparison group on that measure. For studies            measures within studies, yielding a single posttreatment
using between-groups designs, effect sizes were calculated         effect size for each study. Similarly, effect sizes for all de-
with the following formula: d = (Mt – Mc)/SDp, in which           pendent measures completed at all follow-up intervals were
Mt = the mean of the treatment group on a specific mea-            averaged, yielding a single follow-up effect size for each

MINDFULNESS TRAINING         •   BAER                                                                                      131
Table 1. Empirical studies of effects of mindfulness-based interventions

                                                                                Type                Mean        %      Research               Treatment    Control    Rand   %     Dependent                 d      d Foll
                                            Study                N              Participant         Age         Male   Design                 Group        Group      Assn   Att   Variables                 Post   (Months)

                                            Kabat-Zinn            51            Chronic             46          35     Pre-post               MBSR         None       No     12    Pain Rtg Index            0.70   0.35
                                            (1982)                              pain patients                                                                                      Body Parts PA                    (2–7)
                                                                                                                                                                                   Interfere rtg
                                                                                                                                                                                   MSCL
                                                                                                                                                                                   POMS total
                                                                                                                                                                                   SCL-90-R GSI
                                            Kabat-Zinn            90            Chronic             44          33     Pre-post               MBSR         None       No     15    Same as above             0.36   0.51
                                            et al. (1985)                       pain patientsa                                                                                                                      (2–15)
                                            Part 1
                                            Kabat-Zinn            42            Chronic             48          24     Between group          MBSR         TAU        No     —     Same as above             0.26
                                            et al. (1985)                       pain patientsb                                                (n = 21)     (n = 21)
                                            Part 2

CLINICAL PSYCHOLOGY: SCIENCE AND PRACTICE
                                            Kabat-Zinn            30–142        Chronic             —           31     Series of follow-ups   MBSR         None       No     —     Pain Rtg Index            —      0.53
                                            et al. (1987)                       pain patientsc                         2.5–48 months                                               Body Parts PA                    (6–48)

  •
                                                                                                                       post-MBSR                                                   MSCL
                                                                                                                                                                                   SCL-90-R GSI
                                            Randolph              78            Chronic             50          31     Pre-post               MBSR         None       No     —     Pain rtgs                 0.15   0.08
                                            et al. (1999)                       pain patients                                                                                      BSI-GSI                          (2–12)
                                                                                                                                                                                   POMS total
                                                                                                                                                                                   Pain beliefs
                                            Kabat-Zinn            22            Anxiety             38          23     Pre-post               MBSR         None       No      8    Hamilton Anx              0.88   1.35
                                            et al. (1992)                       patients                                                                                           Hanilton Dep                     (3)
                                                                                                                                                                                   BDI; BAI; FSS
                                                                                                                                                                                   Mobility Inv
                                            Miller et al.         18            Anxiety             —           —      3-year follow-up       MBSR         None       No     —     Hamilton Anx              —      1.10
                                            (1995)                              patients                               (patients from                                              Hamilton Dep                     (36)

V10 N2, SUMMER 2003
                                                                                from Kabat-                            Kabat-Zinn et al.,                                          BDI; BAI; FSS
                                                                                Zinn et al.,                           1992)                                                       Mobility Inv
                                                                                1992)
                                            Kristeller &          18            Binge eating        46           0     Pre-post               Variant of   None       No     14    Binge freq                1.65   —
                                            Hallett (1999)                      disorder                                                      MBSR                                 Binge Eat Sc
                                                                                                                                                                                   BDI; BAI
                                                                                                                                                                                   Eating rtgs
                                            Teasdale             132            Remitted            44          24     Between group          MBCT         TAU        Yes    17    MDD relapse               0.60   —
                                            et al. (2000)                       MDD after                                                     (n = 63)     (n = 69)
                                                                                medicn tx
                                            Williams              41            Remitted            43          27     Between group          MBCT         TAU        Yes    —     Autobiog                  0.71   —
                                            et al. (2000)                       MDD after                                                     (n = 21)     (n = 20)                Memory test
                                                                                medicn txd
                                            Kaplan                59            Fibro-myalgia       46          10     Pre-post               Variant of   None       No     23    VAS’s pain, sleep, etc.   —e
                                            et al. (1993)                       patients                                                      MBSR                                 MSCL
                                                                                                                                                                                   SCL-90-R GSI
                                                                                                                                                                                   Coping Strat Q
                                                                                                                                                                                   Fibro Impact Q

132
Goldenberg           121           Fibro-myalgia        46          7         Between group             MBSR              Wait list       No          9      Vas’s pain, sleep, etc.     0.42      —
                       et al. (1994)                      patients                                                             (n = 79)          (n = 18);                          Fibro Impact Q
                                                                                                                                                 declined                           SCL-90-R GSI
                                                                                                                                                 MBSR
                                                                                                                                                 (n = 24)
                       Kabat-Zinn            37           Psoriasis            43         46         Between group             Mindfulness       TAU             Yes        —       Days to clearing            0.63      —
                       et al. (1998)                      patients                                                             tapes during                                         of psoriasis
                                                          receiving                                                            light tx

MINDFULNESS TRAINING
                                                          light therapy

 •
                       Speca et al.          90           Cancer               51         19         Between group             Variant of        Wait list       Yes        13      POMS                        0.60
                       (2000)                             patients                                                             MBSR                                                 SOSI
                       Carlson et al.        54           Cancerf              51         19         6-month follow-           Variant of        None            No         —       POMS                        —         NS
                       (2000)                             patients                                   up in Speca et al.        MBSR                                                 SOSI                                  (6)

BAER
                                                                                                     (2000)
                       Kutz et al.           20           Long-term            38         —          Pre-post                  MBSR              None            No         —       SCL-90-R                    0.72      0.52
                       (1985)                             dynamic                                                                                                                   POMS                                  (6)
                                                          therapy clients                                                                                                           Interfere rtg
                                                                                                                                                                                    Thera rtgs
                       Roth & Creasor        86           Outpatients          43         16         Pre-post                  MBSR              None            NO         40      SCL-90-R                    0.69      —
                       (1997)                             inner city                                                                                                                BAI
                                                          Latino/Latina                                                                                                             MSCL
                                                          Mixed dx                                                                                                                  Self-esteem
                       Reibel et al.        121           Medical              47         29         Pre-post                  MBSR              None            No         11      SF-36                       0.56      NS
                       (2001)                             patients,                                                                                                                 MSCL                                  (12)
                                                          Mixed dx                                                                                                                  SCL-90-R GSI
                       Massion et al.        16           Healthy              42          0         Between group             Regular           Non-            No         —       Melatonin metabolite        0.98      —
                       (1995)                             women                                                                mediators         mediators                          in urine
                       Astin (1997)          19           College              —           5         Between group             MBSR              Wait list       Yes        14      SCL-90-R GSI                1.51      —
                                                          students                                                             (n = 12)          (n = 7)                            Control Inv
                                                                                                                                                                                    INSPIRIT
                       Shapiro et al.        73           Premed and           —          44         Between group             MBSR              Wait list       Yes         3      SCL-90-R GSI                0.50      —
                       (1998)                             med students                                                                                                              SCL-90-R Dep
                                                                                                                                                                                    STAI
                                                                                                                                                                                    Empathy rtg
                                                                                                                                                                                    INSPIRIT
                       Williams              75           Community            43         28         Between group             MBSR              Wait list       Yes        15      DSI                         0.67      0.67
                       et al. (2001)                      volunteers                                                                             and info                           SCL-90-R GSI                          (3)
                                                                                                                                                                                    MSCL

                       Notes. rand assn = random assignment; att = attrition from treatment group; post = posttreatment; foll = follow-up; MDD = major depressive disorder; medicn = medication; tx = treatment; dx = diagnosis;
                       MBSR = Mindfulness-Based Stress Reduction; MBCT = Mindfulness-Based Cognitive Therapy; TAU = Treatment As Usual; rtg = rating; PA = problem assessment; MSCL = Medical Symptom Checklist;
                       POMS = Profile of Mood States; SCL-90-R = Symptom Checklist-90 Revised; GSI = General Severity Index; Anx = anxiety; Dep = depression; BDI = Beck Depression Inventory; BAI = Beck Anxiety Inven-
                       tory; FSS = Fear Survey Schedule; Inv = inventory; Autobiog = autobiographical; VAS = visual analog scale; Strat = strategies; Q = questionnaire; fibro = fibromyalgia; SOSI = Symptoms of Stress Inventory;
                       SF-36 = Short Form 36; INSPIRIT = Index of Core Spiritual Experiences; STAI = State-Trait Anxiety Inventory; DSI = Daily Stress Inventory; NS = no significant difference between posttest and follow-up.
                       a
                         Includes 51 patients from Kabat-Zinn (1982).
                       b
                         Includes 21 patients from Part 1 of this study.
                       c
                         Includes patients from Kabat-Zinn (1982) and Kabat-Zinn et al. (1985).
                       d
                          Subset of patients in Teasdale et al. (2000).
                       e
                         Insufficient data to calculate.
                       f
                        Subset of patients in Speca et al. (2000).

133
study that reported follow-up data. Mean posttreatment         Axis I Disorders. Kabat-Zinn et al. (1992) examined a
and follow-up effect sizes are presented in the final two        sample of 22 patients with generalized anxiety and panic
columns of Table 1.                                            disorders, and found significant improvements in several
    No studies of DBT, ACT, or RP were included, be-           measures of anxiety and depression, both at posttreatment
cause none were found that examined the mindfulness            and at 3-month follow-up. A no-treatment control group
component independently of the behavior change strate-         was not included. Miller, Fletcher, and Kabat-Zinn (1995)
gies also included in these treatment approaches. Thus,        reported a 3-year follow-up of the same participants and
although empirical studies support the efficacy of these         found that treatment gains had been maintained.
treatments (Curry, Marlatt, Gordon, & Baer, 1988; Ito,             Kristeller and Hallett (1999) examined the effects of
Donovan, & Hall, 1988; Koons et al., 2001; Linehan, Arm-       MBSR on binge eating disorder. In a pre-post design with
strong, Suarez, Allmon, & Heard, 1991; Linehan, Heard, &       no control group, 18 female patients showed statistically
Armstrong, 1993; Linehan, Tutek, Heard, & Armstrong,           significant improvements in several measures of eating and
1994; Strosahl, Hayes, Bergan, & Romano, 1998; Zettle &        mood.
Raines, 1989), the relative contribution of mindfulness            Teasdale et al. (2000) examined the effects of MBCT on
training to these treatment effects has not been investi-       rates of depressive relapse in a large sample of patients
gated. In contrast, studies of MBCT were included, be-         whose major depressive disorder (MDD) had remitted af-
cause mindfulness training appears to be the central focus     ter treatment with medication. All participants had dis-
of this approach, although some cognitive techniques have      continued their medications at least 12 weeks before the
been incorporated.                                             study began. Patients were randomly assigned to either
                                                               MBCT (8-week manualized group treatment) or TAU and
General Findings                                               then followed for 1 year. For patients with three or more
Chronic Pain. Four studies have examined the effects of         previous depressive episodes, results showed much lower
MBSR on patients with chronic pain. The first study             relapse rates for MBCT patients (37% of patients relapsed)
(Kabat-Zinn, 1982) describes pre-post data for 51 patients.    than for the TAU group (66% of patients relapsed) during
The second study (Kabat-Zinn, Lipworth, & Burney,              the 1-year follow-up period. However, relapse rates for
1985), has two parts. Part 1 presents pre-post data for a      the MBCT and TAU groups did not differ for patients with
sample of 90 patients, including the 51 patients from          only one or two previous episodes.
Kabat-Zinn (1982). In Part 2, 21 of these 90 patients are          Using a subset of the participants from Teasdale et al.
compared to 21 other pain patients who had received TAU        (2000), J. M. G. Williams et al. (2000) found that those
in the pain clinic but had not participated in MBSR. Parts     who had completed MBCT produced fewer general and
1 and 2 of this study are entered separately in Table 1. The   more specific memories when asked to recall events from
third study (Kabat-Zinn, Lipworth, Burney, & Sellers,          their pasts in response to cue words. The authors speculate
1987) is an extensive series of follow-up evaluations of       that mindfulness training may modify the overgeneral
chronic pain patients who had completed MBSR over the          autobiographical memory believed to be characteristic of
preceding several years, including patients in the previous    individuals with depression (Kuyken & Brewin, 1995).
two studies. Thus, the first four entries in Table 1 are de-
rived from three published articles with overlapping par-      Other Medical Disorders. Two studies have investigated
ticipant samples. Finally, Randolph, Caldera, Tacone, and      effects of MBSR on fibromyalgia. Both reported improve-
Greak (1999) investigated the effects of MBSR in an in-         ments in a variety of symptoms. In a study of psoriasis pa-
dependent sample of 78 chronic pain patients.                  tients, Kabat-Zinn et al. (1998) found that patients who
    In general, findings for chronic pain patients show sta-    listened to mindfulness audiotapes during individual light-
tistically significant improvements in ratings of pain, other   therapy sessions showed quicker clearing of their skin
medical symptoms, and general psychological symptoms.          (Mdn = 65 days) than did patients who received light ther-
Many of these changes were maintained at follow-up eval-       apy alone (Mdn = 97 days). Speca, Carlson, Goodey, and
uations. Most of these comparisons used pre-post designs       Angen (2000) examined the effects of MBSR in a group
with no control group.                                         of cancer patients and reported significant reductions in

CLINICAL PSYCHOLOGY: SCIENCE AND PRACTICE           •   V10 N2, SUMMER 2003                                           134
mood disturbance and stress levels. Carlson, Ursuliak,         effect sizes in this calculation, the effect sizes obtained from
Goodey, Angen, and Speca (2001) reported that these            Kabat-Zinn (1982) and Parts 1 and 2 of Kabat-Zinn et al.
changes were maintained at 6-month follow-up.                  (1985) first were averaged, because these comparisons have
                                                               overlapping participant samples. Similarly, the mean effect
Mixed Clinical Populations. Kutz et al. (1985) studied a       sizes obtained for Teasdale et al. (2000) and J. M. G.
sample of long-term psychodynamic therapy patients with        Williams et al. (2000) were averaged, because these two
diagnoses including anxiety and obsessive neuroses, and        studies also have overlapping participant samples. After
narcissistic and borderline personality disorders. They        these preliminary calculations, 15 independent posttreat-
completed a 10-week MBSR program while continuing              ment mean effect sizes, each from a separate sample, were
with their individual psychotherapy and showed statisti-       available for analysis. Their mean was 0.74 (SD = 0.39).
cally significant improvements in a variety of self- and        When each of these 15 effect sizes was weighted by sample
therapist-rated symptoms. Roth and Creasor (1997) stud-        size, overall mean effect size was 0.59.
ied outpatients from a low-income, primarily Latino pop-
ulation attending an inner city health clinic and showed       Mean Effect Size at Follow-Up
statistically significant improvements on several measures of   Follow-up data were reported less often. Effect sizes at
medical and psychological functioning. Reibel, Greeson,        follow-up ranged from 0.08 to 1.35. Before an overall
Brainard, and Rosenzweig (2001) studied medical patients       mean of these effect sizes was calculated, mean effect sizes
with a variety of medical and psychiatric diagnoses and        obtained from studies with overlapping participant samples
found significant improvements in medical and psycho-           were averaged. The overall mean of these independent
logical symptoms. None of these studies used control           follow-up effect sizes was 0.59 (SD = 0.41).
groups.                                                            Cohen (1977) has described effect sizes of d = 0.2, d =
                                                               0.5, and d = 0.8 as small, medium, and large, respectively.
Nonclinical Populations. Massion, Teas, Hebert, Wert-          Thus, on the average, the literature reviewed here suggests
heimer, and Kabat-Zinn (1995) analyzed urine levels of a       that mindfulness-based interventions have yielded at least
melatonin metabolite in two groups of women. Levels            medium-sized effects, with some effect sizes falling within
were significantly higher in women previously trained in        the large range. Many of the effect sizes calculated for these
MBSR who continued to meditate regularly than in wo-           studies are probably conservative, because several studies
men who had never been trained and did not meditate.           did not present means, standard deviations, or t values,
The authors cite previous findings suggesting that mela-        making it necessary to calculate d from the p value. In many
tonin level may be related to immune function (Bartsch et      cases exact p values were not reported. Instead, for ex-
al., 1992;Guerrero & Reiter, 1992), and suggest that mind-     ample, a p value between .01 and .05 might have been re-
fulness meditation may influence health status through its      ported as .05, which was then used to compute d. Larger
effects on melatonin. Astin (1997) and Shapiro et al. (1998)    ps yield smaller ds. In addition, when findings were re-
studied student populations who completed group MBSR,          ported only as nonsignificant, effect sizes of zero were
reporting significant effects on psychological symptoms,         recorded. If means, SDs, or t values had been reported in
empathy ratings, and spiritual experiences. Both of these      these cases, the calculated effect size might have been larger
studies used waiting-list control groups. Williams, Kolar,     than zero.
Reger, and Pearson (2001) studied community volunteers
who completed MBSR to reduce their stress levels, re-          Relationships Between Mean Effect Size at Posttreatment and
porting significant improvements in medical and psycho-         Study Characteristics
logical symptoms.                                              Relationships between mean effect sizes at posttreatment
                                                               and selected methodological variables can be seen in Table
Mean Effect Size at Posttreatment                              2. (Follow-up effect sizes are not included in this table.)
Posttreatment effect sizes ranged from 0.15 to 1.65. An         The small number of studies available and the noninde-
overall mean of these effect sizes, collapsed across studies,   pendence of some of the effect sizes make statistical analy-
was calculated. In order to include only independent mean      ses of these differences impractical. Thus, these findings

MINDFULNESS TRAINING         •   BAER                                                                                    135
ies reported means, SDs, or t values, effect sizes for these
Table 2. Mean effect size at posttreatment and methodological variables
                                                                              studies were calculated from p values. It is possible that
Variable                                N            Mean d             SD    mean effect size for chronic pain patients might have been
                                                                              larger if these studies had provided additional data.
Research design
  Pre-post                              8              0.71            0.44
  Between group                         10             0.69            0.34
Random assignment
                                                                              Clinical Significance of Findings
(between group)                                                               The clinical significance of the changes reported in these
  Yes                                   7              0.75            0.34
  No                                    3              0.55            0.38   studies is difficult to assess. Several studies reported only
Type of control group                                                         raw scores on dependent measures, whereas others re-
  Wait list                             5              0.74            0.44
  TAU                                   4              0.55            0.20   ported percentage change in scores or the statistical signi-
Participant population
  Chronic pain                          4              0.37            0.24   ficance of the change in scores. In these cases the severity
  Other Axis Ia                         4              0.96            0.47   of participants’ problems before treatment, or their prox-
  Medicalb                              4              0.55            0.09
  Nonclinicalc                          4              0.92            0.44   imity to the normal range of functioning afterwards, can-
Dependent measure
  Pain                                  17             0.31            0.30
                                                                              not readily be determined.
  Other medical (self-rated)d           11             0.44            0.26       In order to assess the clinical significance of some of
  Anxiety                               8              0.70            0.41
  Depression                             5             0.86            0.30   the findings reviewed here, reported raw scores for more
  Stress                                2              0.63            0.02   frequently used dependent measures were converted to
  Global psychologicale                 18             0.64            0.42
  Objective medicalf                     2             0.80            0.25   T-score equivalents or ranges of functioning, with use of
Method of Calculating d
  Using Ms and SDs, or t                10             0.87            0.40   the instruments’ published manuals or profile sheets. For
  Using p                                8             0.48            0.22   example, several studies reported pre- and posttreatment
a
  Includes anxiety, depression, and binge eating.
                                                                              raw scores for the Global Severity Index (GSI) of the
b
  Includes fibromyalgia, psoriasis, and cancer.                               Symptom Checklist 90-Revised (SCL-90-R) (Derogatis,
c
  Includes students and nonclinical volunteers.
d
  Includes fatigue and sleep ratings, and medical symptom checklist.          1983). These scores were converted to T scores (T-score
e
  Includes POMS total mood disturbance, SCL-90-R GSI.                         equivalents for males and females were averaged) and then
f
 Includes urine and skin analysis.
                                                                              averaged across studies. This procedure yielded a mean
                                                                              pretest T score for the GSI of 67, with a mean posttest T
should be interpreted cautiously, as differences may not be                    score of 60. Because T scores have a mean of 50 and a SD
significant. Mean effect sizes were similar for studies using                   of 10, this finding suggests that patients scored nearly 2 SDs
pre-post and between-groups designs. Mean effect size was                      above the mean before treatment and 1 SD above the mean
somewhat larger when participants had been randomly as-                       after treatment. Several studies using the GSI could not be
signed to groups. Studies using waiting-list control groups                   included in this procedure because they did not report
yielded slightly larger effect sizes than those using TAU.                     scores, instead reporting only percentage decrease in scores,
When organized by type of participant, mean effect sizes                       or the statistical significance of the change in scores.
appear somewhat larger for comparisons using nonclinical                          Similar procedures were followed for the Beck Depres-
populations or patients with selected Axis I problems than                    sion Inventory (BDI) and the Beck Anxiety Inventory
for those with chronic pain or medical problems. When                         (BAI) with use of ranges of functioning described in the
organized by the type of dependent variable, mean effect                       manuals (Beck & Steer, 1987, 1993). On the BDI raw
sizes ranged from 0.31 for pain measures to 0.86 for mea-                     scores of 0–9 are considered asymptomatic, whereas scores
sures of depression.                                                          of 10–18 indicate mild to moderate depression. For two
    Finally, effect sizes derived from means and SDs or t val-                 studies reporting BDI scores, the mean pretreatment score
ues were somewhat larger, on average (0.87) than those                        was 16.82 (mild-moderate), and the mean posttreatment
derived from p values (0.48). This finding illustrates the im-                 score was 8.64 (asymptomatic). For two studies reporting
portance of including means, standard deviations, and t                       BAI scores, the mean pretreatment score was 20.43 (mod-
values in future research. Given the small number of avail-                   erate) and the mean posttreatment score was 7.94 (minimal
able studies, examination of interactions between method                      to mild).
of calculating d and other methodological variables is not                        These indications of clinical significance must be con-
feasible. However, because none of the chronic pain stud-                     sidered tentative, because they are based on very few stud-

CLINICAL PSYCHOLOGY: SCIENCE AND PRACTICE                         •    V10 N2, SUMMER 2003                                             136
ies, some of which used uncontrolled pre-post designs.          scores (r = .66) and in BDI scores (r = .59). Astin (1997),
However, they suggest that mindfulness training, on aver-       in a sample of college students, reported that participants
age, may bring participants with mild to moderate psy-          practiced meditation for an average of 30 min per day, 3.5
chological distress into or close to the normal range.          days per week. Reported practice time and improvement
                                                                on the GSI of the SCL-90-R were not significantly corre-
Attrition, Adherence, and Maintenance of Mindfulness Practice   lated. Reibel et al. (2001) reported that 90% of their mixed
Thirteen studies reported both the number of individuals        sample of medical patients practiced three times per week
who agreed to participate in mindfulness training and the       or more and 57% practiced nearly every day, most for 15–
number who completed it. Program completion usually             30 min each time.
was defined as attendance at a minimum number of ses-                Four studies reported the extent to which participants
sions, or was undefined. Percentage of enrolled partici-         trained in mindfulness skills continued to practice these
pants who completed treatment ranged from 60 to 97,             skills after treatment had ended. In a series of follow-up
with a mean of 85% (SD = 8.91). The lowest completion           studies of former MBSR patients, Kabat-Zinn et al. (1987)
rate (60%) was noted by Roth & Creasor (1997), who stud-        noted that 75% of former patients reported that they still
ied an inner city health clinic population. The highest com-    practiced meditation (averaged across follow-up intervals
pletion rate (97%) was reported by Shapiro et al. (1998),       of 6–48 months). Of these patients, 43% meditated regu-
whose participants were premedical and medical students.        larly (≥ three times weekly, ≥ 15 min each time), whereas
    The most extensive analysis of program completion was       19% meditated sporadically (one or two times weekly, ≥ 15
provided by Kabat-Zinn and Chapman-Waldrop (1988),              min each time, or ≥ three times weekly, ≤ 15 min each
who reported completion rates for the MBSR program (at          time), and 38% were classified as marginal meditators
that time known as the Stress Reduction and Relaxation          (< one time weekly for any length of time, or < three
Program) at the University of Massachusetts Medical Cen-        times weekly, < 15 min each time). Practice of yoga two
ter. (This study is not included in Table 1, because it did     or more times per week was reported by 31% of respon-
not examine treatment effects.) During the 2-year period         dents, and 49% reported using awareness of breathing in
examined (1982–1984), 1,155 patients were referred to           daily life often.
the program, mostly by their physicians. Of these patients,         Kabat-Zinn et al. (1992), at 3-month follow-up of 22
75% completed an intake interview, and 90% of those             patients with anxiety disorders, found that 84% reported
interviewed enrolled in the program. Of the 784 patients        practicing meditation or yoga three or more times per
who enrolled, 76% completed the program, whereas 15%            week, for 15–45 min each time. Mindfulness of breathing
dropped out after beginning and 9% never attended a ses-        in daily life was practiced by 95% (77% often and 18% some-
sion. Regression analyses showed that patients with stress-     times). Miller et al (1995) contacted 18 of these patients for
related problems (hypertension, anxiety, sleep disorders,       a 3-year follow-up evaluation and reported that 10 (56%)
etc.) were significantly more likely to complete the pro-        still practiced meditation: 4 regularly, 3 sporadically, and 3
gram than those with chronic pain complaints (lower back,       marginally (as defined above). Sixteen of 18 (89%) reported
headache, etc.). Completers also had somewhat higher            that they used awareness of breathing in daily life (4 often,
pretreatment scores than noncompleters on the GSI and           11 sometimes, and 1 rarely).
the Obsessive-Compulsive (OC) scales of the SCL-90-R.               K. A. Williams et al. (2001), in a sample of community
Within the chronic pain group, women were slightly more         volunteers self-identified as “stressed out,” reported that at
likely than men to complete the program.                        3-month follow-up 81% of MBSR participants were prac-
    Only three studies reported the extent to which partic-     ticing either meditation, yoga, or awareness of breathing in
ipants completed their assigned home practice during the        daily life.
course of the mindfulness intervention. Kristeller and Hal-
lett (1999), in a sample of women with binge eating disor-      Patients’ Reactions to Treatment
der, noted that participants reported engaging in a mean of     In their follow-up study of former MBSR patients, Kabat-
15.82 hr of meditation (SD = 3.15) across the 6-week in-        Zinn et al (1987) found that the majority of those who
tervention program. Reported practice time was signifi-          considered themselves improved since completing MBSR
cantly correlated with improvements in Binge Eating Scale       attributed 50–100% of their improvement to the MSBR

MINDFULNESS TRAINING         •   BAER                                                                                     137
You can also read